针对低可观测目标的检测和跟踪问题,建立了检测前跟踪的目标运动模型和量测数据模型,研究了基于粒子滤波的检测前跟踪算法。为提高基本算法的运算效率和精度,仿真分析了目标影响分辨单元个数、重采样方法对滤波精度的影响,对实际运用中参数的设定、重采样方法的选取有一定的指导意义;在粒子先验分布上,仿真验证了低门限的方法优于均匀分布的方法,它能够充分利用当前时刻的量测信息,从而提高滤波器的联合检测和跟踪性能。

To solve the problem of detection and tracking low-observable target, the target state transition model and measurement model are constructed, and the particle filter based track before detect (PF-TBD) algorithm is studied. To improve the accuracy and efficiency of basic PF-TBD algorithm, the effect of the number of target spreading cells and the resampling method are analyzed by simulation, which is of significance in application. Considering the prior distribution of particles, this paper shows that the lowthreshold method is superior than generating particles uniformly in all the measurement plane, because it can make full use of the measurement information and improve the joint detection and tracking performance of the filter.

TN957

安徽省自然科学基金资助项目(1708085QF149)